Mean-variance optimization under affine GARCH: A utility-based solution

Affine GARCH models have recently been explored in the context of portfolio optimization, although in a quite narrow setting in terms of utility functions and risk aversion. This work notably extends existing results, accommodating a richer class of objective functions for a large family of GARCH models. In particular, our approach allows for connections to constant proportion portfolio insurance (CPPI) and mean-variance portfolio strategies. We explore the latter numerically based on S&P 500 market data, revealing that a GARCH model clearly outperforms a homoscedastic variant in terms of the efficient frontier.

Automated summary

This study extends the application of Affine GARCH models in the field of portfolio optimization, allowing for a richer class of objective functions. Numerical experiments based on S&P 500 market data reveal that the GARCH model outperforms a homoscedastic variant in terms of the efficient frontier.